1. Field of the Invention
The present invention relates to a speech apparatus for transmitting voice signals in which ambient noise at a wired or wireless transmitter location is reduced so that a high quality voice signal is transmitted.
2. Description of the Related Art
FIG. 8 is a configuration diagram showing a conventional example of speech apparatus.
In FIG. 8, the reference numeral 91 designates a transmission signal microphone; 92, a transmitter side output terminal; 93, a voice detector for detecting transmission voice; and 94, an attenuator for attenuating the transmission signal on the basis of the output of the voice detector.
The operation of the apparatus having the aforementioned configuration will be described below.
In addition to a near-end speaker""s voice 101 inputted to the transmission signal microphone 91, near-end ambient noise 102 is also inputted to the microphone.
The voice detector 93 detects the near-end speaker""s voice 101 inputted to the transmission signal microphone 91. When there is no voice 101, that is, when there is only ambient noise 102, the attenuator 94 for attenuating the transmission signal is operated so that the transmission signal is attenuated. When there is voice 101, contrariwise, the voice detector makes the attenuator 94 inoperative so that the transmission voice is not attenuated.
The conventional speech apparatus configured as above has the following problem.
The level of the near-end speaker""s voice varies correspondingly to the near-end ambient noise. This is because a voice amplifying operation is carried out in order to improve the S/N of the near-end speaker""s transmission signal so that the voice level increases as the near-end ambient noise increases and that the voice level decreases as the near-end ambient noise decreases. Accordingly, the reception voice level fluctuates relative to the far-end speaker. There arises a problem that the voice is hard to hear.
The present invention is designed to solve the aforementioned problem, and it is an object of the invention to provide a speech apparatus in which near-end side noise is reduced to reduce the fluctuation of the level of a transmission signal so that the transmission signal easy to hear on the receiver side is transmitted even under noise.
The noise reduced speech apparatus according to the present invention comprises a specific component removing means for removing specific components from input voice, and a level adjusting means for amplifying a voice signal after removal of the specific components to obtain a constant-level output, wherein the voice signal after the level adjustment is transmitted.
Further, the specific component removing means is constituted by a high-pass filter.
Further, the level adjusting means is constituted by a rectifying circuit, an integrating circuit and a voice level adjuster adjusted on the basis of an output of the integrating circuit.
Further, the level adjusting means is constituted by an A/D converter, a voice power calculating portion, a voice power comparator and a digital level adjuster adjusted on the basis of a value after the power calculation and comparison.
Further, the specific component removing means is constituted by a noise collection microphone and a portion equivalent to a subtractor for subtracting noise components collected by the noise collection microphone from the input voice.
Further, a differential microphone with two inputs each which converts input voice into an electric signal is used in place of the specific component removing means, one input of the microphone being used for voice input and noise collection, and the other input being used for noise collection.
Further, a digital noise canceler for performing noise cancellation by frequency analysis of the output of the A/D converter is provided additionally so as to be used as a substitute for the specific component removing means.
Further, a changeover switch for bypassing the level adjusting means is provided-between the specific component removing means and the level adjusting means, and a noise level detecting means is provided so that when the noise level detected by the noise level detecting means is small, the changeover switch is adjusted so that the level adjusting means is bypassed.
Further, a noise level detecting means is provided additionally so that the gain of the level adjusting means is adjusted on the basis of the noise level detected by the noise level detecting means.
The noise reduced speech method according to the present invention comprises the steps of: canceling noise components in an input voice signal; calculating the power of the voice signal after the noise cancellation; comparing the calculated power of the voice signal with a reference voice power level; and adjusting the level of the voice signal after the noise cancellation on the basis of the result of the comparison.